monitoring and mitigating space weather effects for gnss
TRANSCRIPT
![Page 1: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/1.jpg)
Monitoring and mitigating space weather effects
for GNSS applications
Balwinder Arora1, Michael Terkildsen1, German Olivares2
1 Space Weather Services – Australian Bureau of Meteorology2Spire Global Inc
![Page 2: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/2.jpg)
Overview
• Introduction to the Space Weather Services provided at the Bureau of Meteorology
• Space weather impact on precise positioning
• Introduction to PPP-RTK / National Positioning Infrastructure (NPI)
• 3D Tomographic Ionospheric Model
• Model performance / Validation (quiet conditions)
• September 2017 storm
• Comparative PPP-RTK performance through storm conditions
• Summary
![Page 3: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/3.jpg)
• Originally Ionospheric Prediction Service (IPS) 1947-2008.
• 2008 Renamed "Space Weather Services" (SWS) section within Bureau of Meteorology Hazards Prediction Branch.
• Contact details changed [email protected] → [email protected]
Australian Bureau of Meteorology Space Weather Services
www.sws.bom.gov.au
• Australian Space Forecast Centre (ASFC) team consists of
– 4 Senior Space Weather Forecasters (SSWF's).
– 7 Space Weather Forecasters (SWF's)
– Weekly rotation cycle
• Move to 24/7 forecast centre coverage for significant events.
1
![Page 4: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/4.jpg)
1. SWS Overview: Space Weather Network Sensors and Locations
3
![Page 5: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/5.jpg)
1. SWS overview: Online Products and Services
4
![Page 6: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/6.jpg)
Precise positioning and space weather
Space weather impacts vary by system:
Single-frequency positioning: Impacted by absolute
ionospheric delay
Single frequency positioning utilising broadcast model:
impacted by deviation of Klobuchar model from the true
ionosphere.
Differential / augmented positioning: Most
significantly impacted by spatial gradients in
the ionosphere
Network RTK: Impacted by non-linear
gradients and ionospheric variability with
small spatial scales
Positioning using pseudorange → ionospheric error directly impacts positioning algorithm
Positioning using carrier phase → ionospheric error impacts ambiguity resolution / positioning
![Page 7: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/7.jpg)
"Instantaneous, reliable and fit-for-purpose access to positioning and timing information anytime and anywhere across the Australian landscape and its maritime jurisdictions"
National Positioning Infrastructure
![Page 8: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/8.jpg)
3D Tomographic ionospheric model: 3DB-tomion
Figure from M. Hernandez-Pajares, J.M. Juan, J. Sanz, O.L. Colombo, Improving the real-time ionospheric determination from GPS sites at very long distances over the equator, J Geo Res, V. 107, No A10, 1296, doi:10.1029/2001JA009203, (2002).
ሚ𝑆𝑟𝑠 = න
𝑃𝑟
𝑃𝑠
𝑁𝑒𝑑𝑙 + 𝑑𝑟,𝐺𝐹 − 𝑑 ,𝐺𝐹𝑠
𝑁𝑒 ≈ 𝑖𝑗𝑘
𝐼𝐽𝐾
𝑐𝑖𝑗𝑘 𝑡 ∙ 𝜓𝑖 𝜆 ∙ 𝜓𝑗 𝜃 ∙ 𝜓𝑘 ℎ
• ෩𝑺:Biased Slant Total Electron Content.• 𝑵𝒆: Electron Density.• 𝒅:Delay Code Bias.• 𝝍𝒊: Basis function (e.g. splines).• 𝒄𝒊𝒋𝒌 𝒕 : Basis function coefficient
• 𝑰, 𝑱, 𝑲:Number of basis functions in each dimension.
• No thin-shell approach, thus reducing miss-modelling.• TEC is computed by integration of 𝑁𝑒.• Receiver and satellites DCBs do not depend on geometry, whereas
STEC does → geometrically decorrelated from STEC.
![Page 9: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/9.jpg)
• Reference Network (21 GPS receivers; red dots)
• Test sites (28 rovers; yellow dots).
Ionospheric sounding network
![Page 10: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/10.jpg)
How accurate is the 3D B-splines ionospheric model?
![Page 11: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/11.jpg)
• Compares the raw input data with the modelled output
• RMS ranges from 0.02 to 0.07 TECu.
• No geographical trend due to the local-support feature of B-splines.
Ionospheric model performance
Post-fit residuals
Highly accurate model
![Page 12: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/12.jpg)
• RMS for 2D model is ~100 times higher than for 3D models.
• 2D residual RMS is at TECu level (1 TECu ~ 0.1 m) → Cannot support positioning techniques to achieve RMS at cm level in real-time.
• 3D residual RMS is at 10-2 TECulevel (i.e ~ mm) → It might support positioning techniques to achieve RMS at cm level in real -time.
Ionospheric model performance
3D ionospheric model? Why bother?
![Page 13: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/13.jpg)
By how much does the model improve positioning performance?
![Page 14: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/14.jpg)
Ionospheric Model Test Bed
Ultimate validation tool → How well does the model improve GNSS positioning?
![Page 15: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/15.jpg)
• Time To Fix Ambiguity (TTFA)• Time required to resolve each ambiguity to integer• Impacted significantly by the accuracy of the ionospheric model
• Time To Fix Position (TTFP) • Time required for a user to reach a positioning accuracy better than 10cm
• TTFA / TTFP analysed across all sites
• Results analysed in terms of Cumulative Distribution Functions (CDFs) of TTFA and TTFP
Ionospheric Model Test Bed
Performance Metrics:
![Page 16: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/16.jpg)
• Closed loop: Observed STEC at reference sites used as ionospheric corrections in fictitious rover located at reference sites → Baseline network performance
• Ionospheric hybrid model: 3D B-splines ionospheric model with interpolation to rover sites• Float solution: No ionospheric correction provided to rovers.
CDF # epochs / Closed-
loop
# epochs / Ionomodel
# epochs / Float
68% ~22 ~22 ~74
90% ~30 ~34 ~100
• Baselines range from 50 to 230 km.• 1 epoch = 30 ''• 15° elevation mask
Results – Quiet ConditionsTime to Fix Ambiguity (TTFA)
![Page 17: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/17.jpg)
• 15° 𝑒𝑙𝑒𝑣𝑎𝑡𝑖𝑜𝑛 cut off• Baseline ranges from 50 to 230 km.
1. Closed-loop:• H: Uncertainty for 90% of
computed positions is below 10cm in less than 10 epochs.
• V: Uncertainty for 90% of computed positions are below 10cm in less than 40 epochs.
2. Hybrid model:• H: Uncertainty for 90% of
computed positions are below 10cm in less than 20 epochs.
• V: Uncertainty for 90% of computed positions are below 10cm in less than 50 epochs.
Results – Quiet ConditionsTime to Fix Position (TTFP)
![Page 18: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/18.jpg)
What happens during an ionospheric storm?
![Page 19: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/19.jpg)
September 2017 Space Weather Event
X9.3 flareStrong Earthward-directed CME)
AR2673 Ekc/Beta-gamma-delta
Strong positive phase ionospheric storm
(dayside)
Ionograminversion
Tru
e H
eigh
t
Ne
Impact on Precise GNSS??
08 Sep 2017
Weak bottom-side ionospheric signature
(nightside)
Storm1 (dayside)
Storm2 (nightside)
Au
sD
stIn
dex
Geomagnetic Storms (08 Sep 2017)
UT
![Page 20: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/20.jpg)
08 Sep 2017
EDEN
QuietStorm
STEC
[TE
Cu
]
UT
UT
ResultsIonospheric Storm STECs
Aus Dst Index
![Page 21: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/21.jpg)
PPP-RTK performanceQuiet versus Storm (CDFs)
55%
92%
96%
90%
Net
wo
rk
Ro
ver
Ava
ilab
ility Network (reference sites):
Availability* dropped from 96% (quiet) to 90% (storm)
Rover (away from reference sites):Availability* dropped from 92%(quiet) to 55% (storm)
* Availability defined as the % of locations achieving
horizontal positioning better than 10cm accuracy within
20 epochs from initialisation
Quiet
![Page 22: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/22.jpg)
The previous CDFs showed averaged performance over a day…
How does the time evolution of the storm impact the positioning application?
Can the temporal variation in positioning performance help identify an appropriate proxy for space weather impact to
GNSS?
![Page 23: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/23.jpg)
ResultsTime evolution of CDF summary measure (90th percentile TTFP)
Both storm periods (dayside and nightside) degrade positioning performance
Lagged response to the geomagnetic disturbance by ~2hrs
Well correlated with the large scale ionospheric disturbance during the day (summarised by foF2).
Nightside event appears to be related to a topside disturbance (not seen by ionosonde)
Ionospheric disturbance
Geomagnetic disturbance
foF2 (Sydney)
- Dst (Sydney)
TTF
Po
siti
on
<1
0cm
(9
0th
Perc
enti
le)
08 Sep 2017
![Page 24: Monitoring and mitigating space weather effects for GNSS](https://reader033.vdocument.in/reader033/viewer/2022042705/62663122d44959660824de59/html5/thumbnails/24.jpg)
• Quiet conditions: 3D Ionospheric model corrections → TTFA and TTFP similar to closed-loop (baseline/network performance) with >80% availability (of positioning to <10cm within 10 epochs) across the network in the horizontal component.
• Storm conditions: 3D Ionospheric model corrections → TTFF increases across the network around 2 hours after the geomagnetic storm at day time (~03:00-04:00 UT).
• Correlation and delay between DsT and Ambiguity Success Rate.
• No clear correlation between DsT and STEC, TTFF.
• Influence of the plasmasphere on the PPP-RTK platform → lower Ambiguity Success Rate at local night time (~16:00-17:00 UT).
Summary